IP Expo 2009 - Building a Service-Driven Data Center: Making an internal cloud a reality


Published on

In this session we discuss how to turn a virtual infrastructure into a platform for rapid and ITIL compliant business service delivery. We evaluate what
criteria business service delivery and infrastructure leaders should use when evaluating next generation management solutions. We also take a look at how
these products should ease the life of all roles involved in the business service provisioning and management process: business service managers should
be focused on their application not the details of the infrastructure they are running on - as long as their service level objectives are met they are happy.
Infrastructure managers shouldn’t have to be heavily involved each time a new business service is rolled out or changed - as long as they have visibility
into the utilization efficiency of today’s capacity understand the future capacity demand and can leverage physical virtual or cloud resource they are happy.

Published in: Technology, Business
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • Traditional data center management is saddled with a number of inefficient mechanisms across many of the activities necessary to deploy new services. From requirements discovery to provisioning to going live, many of the steps are time consuming, opaque, inflexible, and manual. Part of the promise of deploying a virtual infrastructure was that many of these processes would be accelerated and ultimately would increase flexibility and reduce cost. Much of this promise has been left unfulfilled—even in a virtual infrastructure—due to the difficulty of managing the infrastructure to support the requirements of new applications. Often the process of provisioning a new service involves a number of different teams, each with a different method of interacting. Developing, communicating, and testing requirements often lead to inefficient outcomes. Even in a virtual infrastructure, often new services still end up with dependencies on obtaining and deploying new hardware, often a process that can take 90 days or more. Once a service is launched, the application owner often “loses touch” with the operational performance of the application they are often charged with supporting. Service level metrics and operational capacity is often visible only to the infrastructure manager, and may not be communicated until anomalies occur. The current generation of tools are also clearly designed to force a homogenous environment, essentially driving vendor lock-in, particularly in virtualization management. Finally, the level of automation in the interdepartmental workflows (particularly between application development and infrastructure deployment) is far too low, leading to more errors and service disruptions.
  • The result is a lack of trust and visibility between the infrastructure and application teams. Infrastructure managers question the capacity demanded by the application team, while the application team questions the delivery timeframe. Building structure, visibility, and automation into this process has the potential for making a dramatic positive impact.
  • One of the big impediments to deploying a new service is when that service is dependent on procuring and provisioning new servers, storage, or network devices. Besides just the lead time in receiving, racking, and configuring the devices, a lot of time is consumed by communicating requirements and negotiating the cost of the additional infrastructure. Virtualized infrastructures seek to reduce much of the cost in terms of time and money, but because of service level requirements its often impossible to eliminate this cost entirely. In many cases brand new hardware must still be provisioned even in a virtual infrastructure to support mission critical services.
  • The promise of virtualization is reducing the time, cost, and complexity involved in provisioning infrastructure. Assuming you can avoid procuring and provisioning new hardware capacity, deploying a virtual device should be radically faster! But in that speed, the underlying cost of providing that infrastructure can be lost. There is a danger that the application teams view the provisioning of new servers in a virtual environment as being “free”, which the infrastructure managers know is not really the case, but sometimes have difficulty demonstrating consistently.
  • So how do we get out of this trap? First, having infrastructure managers build a catalog of standardized infrastructure components can provide a lot of structure to a discussion.
  • Are things going to get faster and simpler? Are thing going to be more efficient? There are two related factors that need to be considered here: first, the level of automation that exists and second, the level of self service. As the level of automation goes up, the time to deploy a new business service goes down, this is the obvious benefit. With automation though, also comes consistency, and repeatability. Each time a request is made for the same thing it’s simply a case if rerunning the automation. Automation is not the silver bullet to efficiency though. Efficiency is about the amount of time people spend adding value to the process. So how can you increase the value that that infrastructure team provides? Sure we can remove some of the time spent doing manual repeatable processes through automation, but let’s also remove the time wasted in endless meetings going back and forth over minute custom details and endless questions “can we do this, can we do that”. What we need to do is give controlled direct access to making and fulfilling requests to those that need to consume the infrastructure, the business service managers – in our case Kelly. By letting Kelly serve herself in a controlled manor a number of activities that would typically consume the infrastructure team are removed.
  • Now we are starting to build up a fairly sophisticated blueprint for how to manage an internal cloud. We next need to considered what the under lying infrastructure is, servers, storage, network, OSes, and of course the virtualization layer,, and how you can control vendor lock in. When looking at management systems to build an internal cloud you need to think about your future plans. The more you vertically integrate yourself, the more tie you have with one vendor, which can have its upside and downside. On the upside you can get a more integrated set of functionality within the domain of the vendor. On the down side that level of integration typically does not extend outside the domain of the vendor, and sometime it can be impossible to integrate with other technologies. This is before we consider the implications of one vendor holding all the power over your environment. By looking at management systems that can support a heterogeneous infrastructure, you can be sure that when you want to make an vendor changes to your underlying infrastructure you don’t need to disrupt the efficient operational procedures you have created for your customers, like Kelly. Things start to get really interesting, when you consider possible future plans to make use of external resource like public clouds, why tie yourself to just your internal resources when you could make use of external capacity to run certain services. Additionally, if Kelly has some freedom on selecting infrastructure, and she wants’ to use Amazon EC2 then it’s better if she can do it in a controlled fashion using the same processes that would be used in house, rather than her going rouge with an uncontrolled deployment to the cloud.
  • You see, fundamentally, what BSM is all about is letting you more effectively monitor and manage your IT infrastructure – not as a set of IT components, but rather, as a set of business services – services like order processing, online trading, or simply email. At the highest level, BSM is a suite of solutions that allows you to dynamically link your business services to the underlying applications and the IT infrastructure components that support them. They all work together to help you minimize risk, manage impact, communicate the value or your IT organization. Our modular approach works to help map your IT technology to your applications, and to your business. From there, we formulate a real-time state-driven model of your IT infrastructure. Our BSM service model is an accurate depiction of not only your physical and virtual infrastructure, but we also model your logical services and applications as well. In this way, you have an accurate understanding of not only which components are affected by an IT issue, but also which services as well. This is especially important when prioritizing which issues get fixed first. From this information we also create a Configuration Management Database (or CMDB) which is a trusted source of IT information for both IT and the business. Using a federated approach to link to IT management, configuration and asset data sources and visualize CMDB information through a Web 2.0 based dashboard that incorporates social networking principles ensure greater CMDB usability and accessibility… and if more people use the CMDB, it becomes a more accurate repository. Leveraging the BSM service model and CMDB, we then apply built-in analytics to help determine the impact or root cause of specific IT issues. Finally, through the myMO dashboard, we consolidate and correlate all of the silos of data that you have to deal with today to get to the bottom of an IT issue. Using our BSM dashboard we provide a “single pane of glass" approach to help you rapidly identify and prioritize IT issues in the context of business impact.
  • BSM works by first integrating the silos of information that you use today to manage your IT infrastructure – data that comes from existing IT management, configuration, asset, or business systems you have already deployed. This is a very important concept because our BSM technology is not a rip and replace technology… rather it works with the tools you already have in-place – BSM complements any IT enterprise, no matter which tools you use. How we do it is we leverage the over 70 different adaptors that are built into BSM. Whether you have HP Openview, Tivoli T/EC, BMC Patrol, CA Unicenter, NetIQ, or you are using Sitescope or Topaz performance monitoring products – our BSM solution can integrate them all into one centralized console. Adapters work to gather, normalize, centralize and correlate all of the information about your infrastructure. From there, we automatically build a model of your enterprise and map your physical, virtual, and logical components into one simple service model dashboard. From this dashboard view, it’s easy to see how components and services relate to each other. What’s more users can easily interact with the service model to speed problem resolution when IT issues arise. Using our built-in analytics, IT users can point and click their way to isolate impact and root cause – shortening typical IT outages by 50%, or more.
  • Let’s take another look at BSM dashboard examples. It’s important to know that our dashboards can be tailored to individual roles or customized to suit a number of different needs. Here we’re showing example dashboards for SLA compliance, and telecom services, as well as simple vanilla IT management. But these are just representative examples… most importantly, BSM can be tailored to your own business needs.
  • Moving to a compute cloud provides a new level of flexibility for maximizing resources, reducing time to launch, simplifying provisioning. The freedom provides for simplicity to the application owner to quickly role out new applications. The flexibility and freedom that the cloud provides comes at a cost. A level of control is now turned over to the cloud and more decisions are made automatically. To meet security and auditing needs and to effectively troubleshoot issues, visibility is needed to the intended and actual state of infrastructure, workloads and business services. The service driven data center combines the flexibility of a cloud infrastructure with the visibility that BSM provides to create a truly agile datacenter
  • Building a service-driven data center therefore provides a much clearer separation of duties, where the infrastructure manager manages the available capacity in terms of compute power, storage capacity, and network bandwidth, while the business service architect can focus on development and integrating the applications needed to satisfy the needs of the business. To enable this transformation, the infrastructure manager needs to build out a core catalog of standardized infrastructure components, and a set of standard service levels that can be adhered to for each. Once this is done, its much easier to determine costs for these components and service levels, and roll those costs into services which leverage the individual components. The end result is a way for the infrastructure and application teams to interact which can dramatically improve the efficiency of communication, dramatically reduce the effort on both sides to deploy a service, and dramatically accelerate the pace of innovation, all while lowering costs.
  • IP Expo 2009 - Building a Service-Driven Data Center: Making an internal cloud a reality

    1. 1. The Service-Driven Data Center Jo De Baer Product Manager Systems and Resource Management [email_address]
    2. 2. Agenda <ul><li>Data Center Challenges </li></ul><ul><li>Creating an Internal Cloud </li></ul><ul><li>Business Service Management </li></ul><ul><li>Conclusion </li></ul>
    3. 3. Data Center Challenges
    4. 4. Cost Factors In Today's Enterprises <ul><li>Slow service delivery </li></ul><ul><li>Bringing a new service to life involves multiple teams </li></ul><ul><li>Poor coordination can mean service delivery times of 90+ days </li></ul><ul><li>Lack of transparency </li></ul><ul><li>Poor visibility into the operations of business services makes it hard to guarantee service levels and correctly interpret anomalies </li></ul><ul><li>Inflexibility </li></ul><ul><li>Vendor lock-in prevents you from leveraging best of breed tools </li></ul><ul><li>Too little automation </li></ul><ul><li>Expensive and error prone manual workflows can lead to business service disruptions </li></ul>
    5. 5. The Players Applications Kelly Business Service Manager Jim Infrastructure Manager Business Ron Business Service Owner Network Storage Servers
    6. 6. Deploying a New Application (Traditional) Initial Request Sure, Not a problem! I need a new server. –It’s got to be good
    7. 7. Provisioning a New Server (Traditional) Negotiate the Bill 60 to 90 days or more! Receive the Server Place the Order Provision the Server Thanks! Thanks!
    8. 8. <ul><li>Extremely fast </li></ul><ul><ul><li>0-90 minutes! </li></ul></ul><ul><li>Cost information lost </li></ul><ul><ul><li>Poor financial visibility </li></ul></ul><ul><ul><li>Sacrifices economic incentives </li></ul></ul>Virtual Provisioning Thanks! Thanks! BUT
    9. 9. Creating an Internal Cloud
    10. 10. Standardize Infrastructure Delivery <ul><li>Return cost visibility to the system by delivering standardized infrastructure components </li></ul><ul><li>Infrastructure specification + Service-level agreement </li></ul><ul><ul><li>Account for the cost of providing the requested level of service </li></ul></ul><ul><ul><li>Clearly differentiate between service levels and server capabilities by exposing cost </li></ul></ul><ul><ul><li>Provide an array of different options via catalog </li></ul></ul><ul><li>Standard components and standard SLA’s made available to the business with visible cost structure </li></ul><ul><ul><li>Simplify planning on both sides of the transaction </li></ul></ul>
    11. 11. Speed up, simplify and make the process to deploy a new business service more efficient? Create Efficiency Jim Infrastructure Manager Automation Self-service Kelly Business Service Manager
    12. 12. <ul><li>Support for existing and future infrastructure </li></ul><ul><li>Dealing with heterogeneous environments </li></ul><ul><li>Expanding to infrastructure not under control </li></ul>Cloud Service Cloud Service Heterogeneous Infrastructure Support
    13. 13. DSL Request System CMDB Product Overview Internal Compute Cloud Storage Compute Network Business Service Business Service Business Service Business Service Business Service Business Service Service-Driven Data Center Management Portal Infrastructure Automation External Compute Cloud Business Service Business Service Business Service
    14. 14. Workload Repository Business Service descriptor Business Service Workload Definition Workload Definition Workload Configuration Workload Instance Workload Instance Workload Instance Workload Template Workload Template Workload Template lives in is part of is part of picks and customizes workload template generates Creates workload templates I want a web shop with a service guarantee for 100 concurrent users I'm sick and tired of installing and configuring web shops Christmas is coming up, we're going to have lots of customers Self-Service Provisioning Process sends an email Workload Catalog publishes configure clone Kelly Business Service Manager Ron Business Service Owner Jim Infrastructure Manager
    15. 15. Product walkthrough Kelly - Business Service Manager
    16. 20. Product walkthrough Jim – Infrastructure Manager
    17. 23. Advantages For Infrastructure Managers <ul><li>Reduce rework by creating service templates for common workload types </li></ul><ul><li>Control resource usage by specifying user and workload policies </li></ul><ul><li>Track resource utilization to control expense </li></ul><ul><li>Automatically choose the most efficient data center fabric for each workload in order to achieve service level targets </li></ul>
    18. 24. Advantages For Business Service Managers <ul><li>Simplified creation of business services </li></ul><ul><li>Automatically translate application-specific requirements to workload configurations </li></ul><ul><li>Align and adjust service level targets with real demand </li></ul><ul><li>Reserve resource for business services by calendar or in response to changing demands </li></ul>
    19. 25. Advantages For Business Service Owners <ul><li>Get access to needed business services quicker </li></ul><ul><li>Reduce the cost of providing new business services by streamlining processes </li></ul><ul><li>Increase IT efficiency, transparency </li></ul><ul><li>Recognize IT “value” </li></ul>
    20. 26. Business Service Management
    21. 27. Business Service Management <ul><ul><li>Maps technology to applications and to the business </li></ul></ul><ul><ul><li>Creates a trusted source for IT and the business </li></ul></ul><ul><ul><li>Turns data into powerful intelligence </li></ul></ul><ul><li>Delivers role-based visualization for a diverse community </li></ul><ul><li>Empowers both IT and business management with real-time, role-based dashboards showing health of mission-critical services </li></ul>Business Service Management dynamically links business services to underlying applications, workloads and IT infrastructure components to minimize risk, manage impact and communicate value.
    22. 28. Novell Business Service Management ADM SLM CMDB360 Asset Data Help Desk Business Metrics Configuration Data IT Management Data Discovery Data <ul><li>Integrates & correlates existing IT data into one centralized dashboard </li></ul><ul><li>Automatically models IT, application, & business services </li></ul><ul><li>Intuitive, role-based “service view” speeds problem isolation </li></ul><ul><li>Built-in impact & root-cause analysis shortens resolution time by 50%, or more </li></ul>Centralize IT management data into a service-based dashboard
    23. 29. Role Based Dashboards SLA Compliance IT Management Order Processing Availability Management Call Center Monitoring Telecom Services Fully Customizable – can be tailored to individual needs
    24. 30. Conclusion
    25. 31. The Service-Driven Data Center
    26. 32. The Service-Driven Data Center <ul><li>Lower the barrier to entry for new services </li></ul><ul><li>Reduces deliver time </li></ul><ul><li>Fail fast </li></ul><ul><li>Simplify provisioning </li></ul><ul><li>Maximizes server resources </li></ul><ul><li>Reduce wasted computing infrastructure </li></ul><ul><li>Balance workload demand with resource supply </li></ul><ul><li>Smart growth of infrastructure </li></ul><ul><li>Improves Flexibility </li></ul><ul><li>The free movement of resources promotes effective utilization </li></ul><ul><li>Provides for more continual improvement processes </li></ul>Benefits :
    27. 33. The Service-Driven Data Center <ul><li>Provides clear separation of duties </li></ul><ul><li>Enable infrastructure managers to focus on providing capacity </li></ul><ul><li>Allow service architects to focus on the application </li></ul><ul><li>Builds structure around core capabilities </li></ul><ul><li>Create repeatable processes, capture best practices </li></ul><ul><li>Expose catalog of standardized, reusable infrastructure components </li></ul><ul><li>Shortens innovation lifecycle </li></ul><ul><li>Dramatically reduce effort required to deploy new services </li></ul><ul><li>Accelerate time to market, mitigate risk </li></ul>Benefits :
    28. 34. <ul><li>Novell Business Service Manager™ </li></ul><ul><li>Novell myMO™ Executive Dashboard </li></ul><ul><li>Novell myCMDB™ </li></ul><ul><li>Novell Business Service Level Manager™ </li></ul><ul><li>Novell Business Experience Manager™ </li></ul><ul><li>PlateSpin ® Migrate </li></ul><ul><li>PlateSpin Recon </li></ul><ul><li>PlateSpin Protect </li></ul><ul><li>PlateSpin Forge ® </li></ul><ul><li>PlateSpin Orchestrate </li></ul><ul><li>PlateSpin “Atlantic” (Available 2010H1) </li></ul><ul><li>SUSE ® Linux Enterprise Server </li></ul><ul><li>SUSE Linux Enterprise for System z </li></ul><ul><li>SUSE Linux Enterprise for SAP </li></ul><ul><li>SUSE Linux Enterprise Real Time Extension </li></ul><ul><li>SUSE Studio & JeOS </li></ul>Products Solutions Virtualization and Workload Management Enterprise Linux Servers Business Service Management SDDC Products
    29. 36. Unpublished Work of Novell, Inc. All Rights Reserved. This work is an unpublished work and contains confidential, proprietary, and trade secret information of Novell, Inc. Access to this work is restricted to Novell employees who have a need to know to perform tasks within the scope of their assignments. No part of this work may be practiced, performed, copied, distributed, revised, modified, translated, abridged, condensed, expanded, collected, or adapted without the prior written consent of Novell, Inc. Any use or exploitation of this work without authorization could subject the perpetrator to criminal and civil liability. General Disclaimer This document is not to be construed as a promise by any participating company to develop, deliver, or market a product. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions. Novell, Inc. makes no representations or warranties with respect to the contents of this document, and specifically disclaims any express or implied warranties of merchantability or fitness for any particular purpose. The development, release, and timing of features or functionality described for Novell products remains at the sole discretion of Novell. Further, Novell, Inc. reserves the right to revise this document and to make changes to its content, at any time, without obligation to notify any person or entity of such revisions or changes. All Novell marks referenced in this presentation are trademarks or registered trademarks of Novell, Inc. in the United States and other countries. All third-party trademarks are the property of their respective owners.